Bot Wars Evolved: Orchestrating Competing LLMs in a Counterstrike Against Phone Scams
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866929751390683136 |
|---|---|
| author | Basta, Nardine Atkins, Conor Kaafar, Dali |
| author_facet | Basta, Nardine Atkins, Conor Kaafar, Dali |
| contents | We present "Bot Wars," a framework using Large Language Models (LLMs) scam-baiters to counter phone scams through simulated adversarial dialogues. Our key contribution is a formal foundation for strategy emergence through chain-of-thought reasoning without explicit optimization. Through a novel two-layer prompt architecture, our framework enables LLMs to craft demographically authentic victim personas while maintaining strategic coherence. We evaluate our approach using a dataset of 3,200 scam dialogues validated against 179 hours of human scam-baiting interactions, demonstrating its effectiveness in capturing complex adversarial dynamics. Our systematic evaluation through cognitive, quantitative, and content-specific metrics shows that GPT-4 excels in dialogue naturalness and persona authenticity, while Deepseek demonstrates superior engagement sustainability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_07036 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Bot Wars Evolved: Orchestrating Competing LLMs in a Counterstrike Against Phone Scams Basta, Nardine Atkins, Conor Kaafar, Dali Computation and Language Artificial Intelligence We present "Bot Wars," a framework using Large Language Models (LLMs) scam-baiters to counter phone scams through simulated adversarial dialogues. Our key contribution is a formal foundation for strategy emergence through chain-of-thought reasoning without explicit optimization. Through a novel two-layer prompt architecture, our framework enables LLMs to craft demographically authentic victim personas while maintaining strategic coherence. We evaluate our approach using a dataset of 3,200 scam dialogues validated against 179 hours of human scam-baiting interactions, demonstrating its effectiveness in capturing complex adversarial dynamics. Our systematic evaluation through cognitive, quantitative, and content-specific metrics shows that GPT-4 excels in dialogue naturalness and persona authenticity, while Deepseek demonstrates superior engagement sustainability. |
| title | Bot Wars Evolved: Orchestrating Competing LLMs in a Counterstrike Against Phone Scams |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2503.07036 |